A method for identifying noisy data

Abstract

The present invention relates to the field of digital signal processing technology, and specifically to a method and device for identifying noise data, comprising: establishing an algorithm in a sound recognition system, so that the device equipped with the system can perform a judgment execution process on the analog audio signal after capturing external sound, thereby eliminating noise. The system of the present invention sets an algorithm internally, so that after the device equipped with the system captures and collects the sound, it identifies and detects the noise part in the sound through a judgment process, so that a large amount of environmental noise interference can be reduced when the audio analog signal is modulated later, and after the audio analog signal captured and collected by the device is eliminated of environmental sound through the system's hierarchical judgment execution strategy, a noise smoothing execution strategy is implemented on it, so that a high-quality digital analog signal storage file can be obtained in the end.

Description

Technical Field

[0001] The present invention relates to the technical field of digital signal processing, and in particular to a method for identifying noise data. Background Art

[0002] Recording high-quality lossless audio in a browser is challenging, especially because the devices running the browser exist in different environments and will generate multiple sounds when recording sound.

[0003] When trying to record audio using a browser, the device will capture the main input sound of the microphone, and at the same time it will capture the surrounding environmental sound, resulting in the presence of multiple sound fragments in the audio analog signal collected by the device. Because the presence of multiple sound fragments will interfere with the main input audio, the recorded audio will be confused by noise. Therefore, it is necessary to determine and eliminate the noise of the collected audio analog signal before the device modulates it to reduce the noise interference in the audio. After the audio analog signal collected by the device is converted into a digital analog signal, there may still be a small amount of noise that cannot be eliminated, which will cause the quality of the obtained audio file to deteriorate. Therefore, it is urgent to design a noise data recognition method to solve the above problem. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for identifying noise data to solve the problem raised in the above-mentioned background technology that the presence of multiple sound clips will interfere with the main input audio, causing the recorded audio to be confused by noise. Therefore, it is necessary to determine and eliminate the noise of the collected audio analog signal before the device modulates it to reduce the noise interference in the audio. After the audio analog signal collected by the device is converted into a digital analog signal, a small amount of noise that cannot be eliminated may still exist therein, which will lead to a decrease in the quality of the obtained audio file.

[0005] To achieve the above-mentioned object, the present invention provides the following technical solution: by establishing an algorithm in a sound recognition system, the device equipped with the system can perform a judgment execution process on the analog audio signal after capturing external sound, thereby eliminating noise. The system algorithm includes a five-level judgment strategy, which includes a first-level judgment, a second-level judgment, a third-level judgment, a fourth-level judgment, and a fifth-level judgment. The first-level judgment is for judging the loudness of the sound, the second-level judgment is for judging the loudness range of the audio analog signal, the third-level judgment is for judging the loudness classification of the sound, the fourth-level judgment is for judging the existence of the permanence of the sound, and the fifth-level judgment is for judging the data smoothness of the audio digital signal. The first-level judgment is for judging the system algorithm for judging the audio analog signal. In the first strategy implementation, the first-level determination is to determine the sound loudness, specifically, to determine whether the lowest loudness sound in the audio analog signal exceeds a specified value. If the first-level determination is effective, a first-level execution command is executed on the audio analog signal. The first-level execution is to reduce the sound loudness value of the audio analog signal, specifically, to reduce the loudness value of the entire captured audio analog signal until the lowest loudness value in the audio analog signal reaches a value specified by the system. The value specified by the system can be set to the minimum limit that a normal person can distinguish sound from. After the first-level execution is completed, the processed audio analog signal enters the second-level determination. If the first-level determination is not effective, the second-level determination is performed on the audio analog signal captured by the device.

[0006] Preferably, the secondary judgment is to make a judgment on the loudness range of the audio analog signal, specifically to judge whether there is an audio segment in the signal that exceeds the hearing tolerance range of a normal person. When the secondary judgment takes effect, the secondary execution command is performed on the audio analog signal. The secondary execution command is to cut off a segment of the audio analog signal, specifically to cut off the audio analog signal segment targeted by the previous judgment. After completion, the secondary execution command will cause the processed audio analog signal to enter the third level judgment. After the secondary judgment fails, the audio analog signal will enter the third level judgment.

[0007] Preferably, the three-level determination is to determine the loudness grading of the sound, specifically to determine whether there is an audio segment with obvious loudness grading in the captured audio analog signal. When the three-level determination fails, the audio analog signal will enter the third-level execution, and the third-level execution is to pulse encode the analog signal. When the three-level determination is effective, the audio analog signal will enter the fourth-level determination.

[0008] Preferably, the four-level judgment is to make a judgment on the constant existence of sound, specifically to determine whether the audio segment that is different from the normal sound in the audio analog signal is in a state of long-term existence. When the four-level judgment takes effect, the audio signal segment will enter the third level execution. When the four-level judgment fails, the audio signal segment will enter the second level execution.

[0009] Preferably, after the third-level execution processes the audio analog signal, the audio analog signal will be converted into an audio digital signal. After the third-level execution completes the strategy implementation, the audio digital signal will enter the fifth-level determination.

[0010] Preferably, the five-level judgment is to judge the data smoothness of the audio digital signal, specifically to judge whether the audio digital signal meets the data fluctuation range set by the system. When the five-level judgment is effective, the audio digital signal will enter the fifth level execution. When the five-level judgment fails, the audio digital signal will enter the fourth level execution.

[0011] Preferably, the fourth level execution is to smooth the noise signal portion determined in the data of the audio digital signal, so that it approaches a state of stable fluctuation with a normal digital audio segment. After completing the processing of the audio digital signal, the fourth level execution will allow the processed audio digital signal to enter the fifth level execution, and the fifth level execution is to store the data of the audio digital signal.

[0012] Compared with the prior art, the present invention has the following beneficial effects:

[0013] The system of the present invention has an algorithm set up internally, so that after the device equipped with the system captures and collects the sound, the audio analog signal converted therefrom will be graded and judged by the system algorithm for the implementation of the execution strategy. Through the judgment process, the noise part in the sound is identified and detected, and then the execution strategy is implemented on it, so that the noise existing in the audio analog signal is gradually eliminated, so that a large amount of environmental noise interference can be reduced when the audio analog signal is modulated later.

[0014] After the audio analog signal captured by the device is eliminated of ambient sound through the system's hierarchical judgment execution strategy, there are still some noises that cannot be distinguished from the main input audio signal. At this time, after the audio analog signal is converted into an audio digital signal through the system algorithm, a noise smoothing execution strategy is implemented on it, so that high-quality digital analog signal storage files can be obtained in the end. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a schematic diagram of the algorithm flow for system hierarchical determination execution strategy of the present invention; DETAILED DESCRIPTION

[0016] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0017] See also Figure 1 , an embodiment provided by the present invention:

[0018] A noise data recognition method includes: establishing an algorithm in a sound recognition system so that a device equipped with the system can perform a judgment execution process on an analog audio signal after capturing external sound, thereby eliminating noise. The system algorithm includes a five-level judgment strategy, which includes a first-level judgment, a second-level judgment, a third-level judgment, a fourth-level judgment, and a fifth-level judgment. The first-level judgment is for judging the loudness of the sound, the second-level judgment is for judging the loudness range of the analog audio signal, the third-level judgment is for judging the loudness classification of the sound, the fourth-level judgment is for judging the existence of the constancy of the sound, and the fifth-level judgment is for judging the data smoothness of the digital audio signal. Through the hierarchical judgment execution strategy set in the system algorithm, hierarchical detection and processing can be performed on the analog audio signal captured by the device after it is captured, so as to minimize the influence of noise in the obtained audio signal, thereby ensuring that the device can record high-quality audio files.

[0019] Furthermore, the first-level determination is the first strategic implementation of the system algorithm on the audio analog signal. The first-level determination is the determination of the sound loudness, specifically whether the lowest loudness sound in the audio analog signal exceeds the limited value. When the first-level determination is effective, the audio analog signal is executed at the first level. The first-level execution is to reduce the sound loudness value of the audio analog signal, specifically to reduce the loudness value of the overall captured audio analog signal until the lowest loudness value in the audio analog signal reaches the value position limited by the system. The value limited by the system can be set to the minimum limit value that a normal person can distinguish from the sound. The first-level execution After completion, the processed audio analog signal will enter the second-level judgment. If the first-level judgment is not effective, the second-level judgment will be performed on the audio analog signal captured by the device. The first-level judgment strategy set by the algorithm can prioritize the loudness value of the recorded audio signal and reduce its overall value to the minimum loudness value set by the system to minimize the maximum loudness value of the overall audio signal from exceeding the hearing tolerance range of a normal person. After the first-level execution strategy is executed and the loudness value of the overall audio signal has been reduced, the second-level judgment is performed to detect whether there are still loudness values ​​in the audio signal that exceed the hearing range of a normal person.

[0020] Furthermore, the second-level judgment is to make a judgment on the loudness range of the audio analog signal, specifically to determine whether there is an audio segment in the signal that exceeds the hearing tolerance range of a normal person. When the second-level judgment takes effect, a second-level execution command is executed on the audio analog signal. The second-level execution command is to cut off segments of the audio analog signal, specifically to cut off the audio analog signal segment targeted by the previous-level judgment. After the second-level execution command is completed, the processed audio analog signal will enter the third-level judgment. After the second-level judgment fails, the audio analog signal will enter the third-level judgment. When the overall loudness value of the audio signal is reduced, the maximum loudness value will be determined through the second-level judgment. If there are still loudness value segments that exceed the hearing tolerance range of a normal person, they will be cut off through the second-level execution strategy, thereby ensuring that there are no sounds with excessive loudness values ​​in the obtained audio signal segments.

[0021] Furthermore, the three-level judgment is to make a judgment on the loudness classification of the sound, specifically to determine whether there is an audio segment with obvious loudness classification in the captured audio analog signal. When the three-level judgment fails, the audio analog signal will enter the three-level execution, and the three-level execution is to pulse encode the analog signal. When the three-level judgment takes effect, the audio analog signal will enter the four-level judgment. The three-level judgment strategy is set to detect whether there is a sound in the sound captured by the device that is significantly different from the main input sound. If this situation does not exist, the audio analog signal can be directly encoded and converted into a digital analog signal. If there is a sound signal that is significantly different from the main input audio signal, it will enter the next level of judgment detection.

[0022] Furthermore, the four-level judgment is to make a judgment on the constant existence of sound, specifically to determine whether the audio segment that is different from the normal sound in the audio analog signal is in a state of long-term existence. When the four-level judgment takes effect, the audio signal segment will enter the third-level execution. When the four-level judgment fails, the audio signal segment will enter the second-level execution. The four-level judgment is set to detect whether the sound signal that is different from the main input signal in the sound signal is noise. When it exists constantly, it can be judged as the ambient sound that is different from the noise in the environment, and it is encoded together with the main input sound signal. When it is not a constantly existing sound, it can be judged as unstable noise in the environment, and it is eliminated through the system's second-level execution strategy, and then continue to enter the third-level judgment detection, thereby completing the subsequent process.

[0023] Furthermore, after the third-level execution processes the audio analog signal, the audio analog signal will be converted into an audio digital signal. After the third-level execution strategy is implemented, the audio digital signal will enter the fifth-level judgment. The setting of the third-level execution is to convert the audio analog signal into an audio analog signal to ensure long-term storage in the later stage.

[0024] Furthermore, the five-level judgment is to judge the data smoothness of the audio digital signal, specifically to judge whether the audio digital signal meets the data fluctuation range set by the system. When the five-level judgment is effective, the audio digital signal will enter the five-level execution. When the five-level judgment fails, the audio digital signal will enter the four-level execution. The setting of the five-level judgment is to perform data detection on the converted audio digital signal. If there is a data point with a larger fluctuation amplitude than the adjacent data, it will enter the four-level execution to eliminate noise.

[0025] Furthermore, the fourth level execution is to smooth the noise signal portion determined in the data of the audio digital signal, so that it approaches a state of stable fluctuation with the normal digital audio segment. After completing the processing of the audio digital signal, the fourth level execution will allow the processed audio digital signal to enter the fifth level execution. The fifth level execution is to store the data of the audio digital signal. The setting of the fourth level execution can perform data averaging processing on the audio digital signal to achieve the smoothing of the overall data signal. In this way, the noise existing in the audio digital signal can be eliminated.

[0026] Working principle: When the system algorithm device equipped with a hierarchical judgment execution strategy records the sound, it captures and collects the sound signal internally and converts it into an audio analog signal. Then the audio analog signal enters the judgment range of the system algorithm. It first passes the first-level judgment of the algorithm to detect whether the lowest loudness value in the sound signal is higher than the standard adjustment value set inside the system. When the lowest loudness value in the signal is higher than the standard adjustment value, the audio analog signal will enter the first-level execution, thereby reducing the loudness value of the overall audio signal to ensure that the captured audio analog signal is processed to the maximum extent. If the minimum loudness value in the audio signal is less than or equal to the standard calibration value, it will directly enter the second-level judgment. The second-level judgment is to ensure that the maximum loudness value is still beyond the normal hearing tolerance range after the minimum sound value is no higher than the standard calibration value. If it is beyond the range, it can be judged as a noise existing in the environment. The sound that exceeds the normal hearing tolerance range will be cut off, and then it will enter the third-level judgment. When the third-level judgment is performed, there is no sound beyond the hearing range in the audio analog signal. The system then determines whether there is a sound segment in the sound signal that is significantly different from the main input audio. If not, it indicates that the obvious noise in the sound signal has been eliminated and can be encoded. If there is a sound with obvious difference, the system will perform a three-level judgment on the different sound signal segment. The three-level judgment is to detect whether the different sound exists for a long time. If it exists for a long time, the system can determine that it is a background sound in the environment. If the sound signal exists discontinuously or intermittently, it can be determined that it is a sudden noise in the environment and can be cut off. Then, the analog signal is encoded and converted into a digital signal. After the form of the sound signal is converted, the final five-level judgment will be performed. This judgment is to detect whether the signal data in the digital signal is overall stable. If so, it can be determined that there is no noise. If not, the adjacent data of the abrupt points are averaged to make the overall data smooth, thereby eliminating the noise and obtaining a high-quality audio storage file.

[0027] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A method for identifying noise data, characterized in that: include: By establishing an algorithm in the sound recognition system, the device equipped with the system can perform a judgment execution process on the analog audio signal after capturing external sounds, thereby eliminating noise. The system algorithm includes a five-level judgment strategy, which includes first-level judgment, second-level judgment, third-level judgment, fourth-level judgment, and fifth-level judgment. The first-level judgment is for judging the loudness of the sound, the second-level judgment is for judging the loudness range of the audio analog signal, the third-level judgment is for judging the loudness classification of the sound, the fourth-level judgment is for judging the existence of the permanence of the sound, and the fifth-level judgment is for judging the data smoothness of the audio digital signal. The first-level judgment is the first strategy implementation of the system algorithm on the audio analog signal. The first-level determination is a determination of sound loudness, specifically, a determination of whether the lowest loudness sound in the audio analog signal exceeds a specified value. If the first-level determination is effective, a first-level execution command is executed on the audio analog signal. The first-level execution command is to reduce the loudness value of the audio analog signal, specifically, to reduce the loudness value of the entire captured audio analog signal until the lowest loudness value in the audio analog signal reaches a value specified by the system. The value specified by the system can be set to the minimum value that a normal person can distinguish from sound. After the first-level execution is completed, the processed audio analog signal enters a second-level determination. If the first-level determination is not effective, a second-level determination command is executed on the audio analog signal captured by the device.

2. The noise data recognition method according to claim 1, characterized in that: The secondary judgment is to make a judgment on the loudness range of the audio analog signal, specifically to determine whether there is an audio segment in the signal that exceeds the hearing tolerance range of a normal person. When the secondary judgment is effective, the audio analog signal is subjected to a secondary execution command. The secondary execution command is to cut off a segment of the audio analog signal, specifically to cut off the audio analog signal segment targeted by the previous level judgment. After completion of the secondary execution command, the processed audio analog signal will enter the third level judgment. After the second level judgment fails, the audio analog signal will enter the third level judgment.

3. The noise data recognition method according to claim 2, characterized in that: The three-level determination is to determine the loudness level of the sound, specifically to determine whether there is an audio segment with obvious loudness level in the captured audio analog signal. When the three-level determination fails, the audio analog signal will enter the third-level execution, which is to pulse encode the analog signal. When the three-level determination is effective, the audio analog signal will enter the fourth-level determination.

4. The noise data identification method according to claim 1, characterized in that: The four-level judgment is to determine the constant existence of sound, specifically to determine whether the audio segment that is different from the normal sound in the audio analog signal is in a state of long-term existence. When the four-level judgment is effective, the audio signal segment will enter the third level execution. When the four-level judgment fails, the audio signal segment will enter the second level execution.

5. The noise data identification method according to claim 1, characterized in that: After the third level execution processes the audio analog signal, the audio analog signal will be converted into an audio digital signal. After the third level execution completes the strategy implementation, the audio digital signal will enter the fifth level judgment.

6. The noise data identification method according to claim 1, characterized in that: The five-level judgment is to judge the data smoothness of the audio digital signal, specifically to judge whether the audio digital signal meets the data fluctuation range set by the system. When the five-level judgment is effective, the audio digital signal will enter the five-level execution. When the five-level judgment fails, the audio digital signal will enter the four-level execution.

7. The noise data identification method according to claim 1, characterized in that: The fourth level execution is to smooth the noise signal portion determined in the data of the audio digital signal so that it approaches a state of stable fluctuation with a normal digital audio segment. After completing the processing of the audio digital signal, the fourth level execution will allow the processed audio digital signal to enter the fifth level execution, and the fifth level execution is to store the data of the audio digital signal.

Images